iEnhancer-RF: Identifying enhancers and their strength by enhanced feature representation using random forest

Enhancers are short DNA regions bound with activators to increase gene transcription over long distances. Hence, they play a crucial role in regulating eukaryotic gene expression. Because enhancers are present in unique genomic regions and have dynamic natures, they are challenging to identify or characterize. Existing experimental methods are time-consuming and expensive, highlighting the need for computational approaches to accelerate the genome-wide identification of enhancers. In this study, the proposed iEnhancer-RF model for an enhancer and its strength was composed of feature representation, feature selection, and predictor construction. The results showed that iEnhancer-RF gives outstanding performance compared to the existing state-of-the-art models. In more concrete terms, iEnhancer-RF improved the independent test accuracy of the enhancer and enhancer strength identification by 2.25% and 10%, respectively. A web server for the iEnhancer-RF is easily accessible at http://nsclbio.jbnu.ac.kr/tools/iEnhancer-RF/.